Message Board

Dear readers, authors and reviewers,you can add a message on this page. We will reply to you as soon as possible!

2021 Volume 43 Issue 6
Article Contents

YANG Hao, WU Jian-rong. Pseudo-Metric Structure and Isometrical Isomorphism in Fuzzy Metric Spaces[J]. Journal of Southwest University Natural Science Edition, 2021, 43(6): 95-100. doi: 10.13718/j.cnki.xdzk.2021.06.013
Citation: YANG Hao, WU Jian-rong. Pseudo-Metric Structure and Isometrical Isomorphism in Fuzzy Metric Spaces[J]. Journal of Southwest University Natural Science Edition, 2021, 43(6): 95-100. doi: 10.13718/j.cnki.xdzk.2021.06.013

Pseudo-Metric Structure and Isometrical Isomorphism in Fuzzy Metric Spaces

More Information
  • Corresponding author: WU Jian-rong
  • Received Date: 17/05/2020
    Available Online: 20/06/2021
  • MSC: O189.13

  • Due to the successful applications in such fields as color image filtering, the research about fuzzy metrics has attracted more and more attention in recent years. While many important results of classical metric spaces are generalized to fuzzy metric spaces, the innovation of research methodsis becoming increasingly important. No doubt, decomposition of fuzzy metrics into a family of classical metrics and establishment of decomposition theorems of fuzzy metrics are of great significance. The existing decomposition theorems, however, are mainly focused on fuzzy metrics with the minimum operator, and thus have great limitations in application. In this paper, the concept of family of star pseudo-metrics is introduced as a generalization of the family of pseudo-metrics. In addition, the sufficient and necessary conditions of isometric isomorphism between a fuzzy metric space and a family of pseudo-metric spaces are given, and the relationship between them is constructed, which provides a new and effective approach to the study of fuzzy metrics in the general sense.
  • 加载中
  • [1] KRAMOSIL I, MICHALEK J. Fuzzy Metrics and Statistical Metric Spaces[J]. Kybernetika, 1975, 11(5): 336-344.

    Google Scholar

    [2] GEORGE A, VEERAMANI P. On Some Results in Fuzzy Metric Spaces[J]. Fuzzy Sets and Systems, 1994, 64(3): 395-399. doi: 10.1016/0165-0114(94)90162-7

    CrossRef Google Scholar

    [3] SHI F G. (L, M)-Fuzzy Metric Spaces[J]. Indian Journal of Mathematics, 2010, 52(2): 231-250.

    Google Scholar

    [4] GAO Y, ZHOU X N. The Relationships Between KM-Fuzzy Quasi-Metric Spaces and the Associated Posets of Formal Balls[J]. Electronic Notes in Theoretical Computer Science, 2017, 333: 17-29. doi: 10.1016/j.entcs.2017.08.003

    CrossRef Google Scholar

    [5] GREGORI V, ROMAGUERA S. On Completion of Fuzzy Metric Spaces[J]. Fuzzy Sets and Systems, 2002, 130(3): 399-404. doi: 10.1016/S0165-0114(02)00115-X

    CrossRef Google Scholar

    [6] GREGORI V, ROMAGUERA S. Characterizing Completable Fuzzy Metric Spaces[J]. Fuzzy Sets and Systems, 2004, 144(3): 411-420. doi: 10.1016/S0165-0114(03)00161-1

    CrossRef Google Scholar

    [7] RANO G, BAG T, SAMANTA K S. Fuzzy Metric Space and Generating Space of Quasi-Metric Family[J]. Ann Fuzzy Math Inform, 2016, 11(2): 183-195.

    Google Scholar

    [8] SANCHEZ I, SANCHIS M. Fuzzy Quasi-Pseudometrics on Algebraic Structures[J]. Fuzzy Sets and Systems, 2018, 330: 79-86. doi: 10.1016/j.fss.2017.05.022

    CrossRef Google Scholar

    [9] 杨洋, 吴健荣. 关于模糊拟度量诱导的双拓扑空间的一些性质[J]. 苏州科技大学学报(自然科学版), 2017, 34(4): 14-19. doi: 10.3969/j.issn.1672-0687.2017.04.003

    CrossRef Google Scholar

    [10] 杨浩, 吴健荣. 模糊度量空间中的有界集[J]. 西南大学学报(自然科学版), 2019, 41(10): 45-50.

    Google Scholar

    [11] GREGORI V, MORILLAS S, SAPENA A. Examples of Fuzzy Metrics and Applications[J]. Fuzzy Sets and Systems, 2011, 170(1): 95-111. doi: 10.1016/j.fss.2010.10.019

    CrossRef Google Scholar

    [12] MORILLAS S, GREGORI V, PERIS-FAJARNES G, et al. A New Vector Median Filter Based on Fuzzy Metrics[C]//Image Analysis and Recognition: Second International Conference. Toronto: Mohamed Kamel Aurélio Campilho, 2005: 81-90.

    Google Scholar

    [13] MORILLAS S, GREGORI V, PERIS-FAJARNES G, et al. A Fast Impulsive Noise Color Image Filter Using Fuzzy Metrics[J]. Real-Time Imaging, 2005, 11(5-6): 417-428. doi: 10.1016/j.rti.2005.06.007

    CrossRef Google Scholar

    [14] ROMAGUERA S, SAPENA A, TIRADO P. The Banach Fixed Point Theorem in Fuzzy Quasi-Metric Spaces with Application to the Domain of Words[J]. Topology and Its Applications, 2007, 154(10): 2196-2203. doi: 10.1016/j.topol.2006.09.018

    CrossRef Google Scholar

    [15] WU J R, JIN Z Y. A Note on Ulam Stability of Some Fuzzy Number-Valued Functional Equations[J]. Fuzzy Sets and Systems, 2019, 375: 191-195. doi: 10.1016/j.fss.2018.10.018

    CrossRef Google Scholar

    [16] 张振荣, 赵凯. 非齐度量测度空间上广义分数次积分算子交换子的有界性[J]. 西南大学学报(自然科学版), 2020, 42(8): 88-96.

    Google Scholar

    [17] 张韩雨, 陈守全. Brinbaum-Sauders分布的极值收敛速度[J]. 西南师范大学学报(自然科学版), 2020, 45(1): 19-24.

    Google Scholar

    [18] SCHWEIZER B, SKLAR A. Statistical Metric Spaces[J]. Pacific J Math, 1960, 10(1): 313-334. doi: 10.2140/pjm.1960.10.313

    CrossRef Google Scholar

  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Article Metrics

Article views(869) PDF downloads(349) Cited by(0)

Access History

Other Articles By Authors

Pseudo-Metric Structure and Isometrical Isomorphism in Fuzzy Metric Spaces

    Corresponding author: WU Jian-rong

Abstract: Due to the successful applications in such fields as color image filtering, the research about fuzzy metrics has attracted more and more attention in recent years. While many important results of classical metric spaces are generalized to fuzzy metric spaces, the innovation of research methodsis becoming increasingly important. No doubt, decomposition of fuzzy metrics into a family of classical metrics and establishment of decomposition theorems of fuzzy metrics are of great significance. The existing decomposition theorems, however, are mainly focused on fuzzy metrics with the minimum operator, and thus have great limitations in application. In this paper, the concept of family of star pseudo-metrics is introduced as a generalization of the family of pseudo-metrics. In addition, the sufficient and necessary conditions of isometric isomorphism between a fuzzy metric space and a family of pseudo-metric spaces are given, and the relationship between them is constructed, which provides a new and effective approach to the study of fuzzy metrics in the general sense.

  • 为描述两点距离的不确定性,文献[1]给出了模糊度量(简称为KM模糊度量)的概念,文献[2]对KM模糊度量进行了改进,提出了现在被称之为GV模糊度量的新概念. 文献[3]对KM模糊度量和GV模糊度量进行了推广,引入了(LM)模糊度量的概念. 到目前为止,许多经典度量空间的重要结果被推广到了模糊度量空间中[4-10],同时,模糊度量已经被广泛地应用在彩色图像处理和算法分析中[11-17]. 为研究模糊度量与分明度量之间的关系,文献[7]给出了伪度量族空间的概念,建立了两个分解定理. 然而正如文献[7]中所指出的,这两个定理成立需要对模糊度量定义中的t-模进行严格的限制.

    本文引入了星伪度量族的概念,利用这一概念,建立了具有一般t-模的模糊度量的分解定理. 此外,在引入模糊度量空间与伪度量族空间等距同构的概念之后,给出了模糊度量空间与伪度量族空间等距同构的充分条件和必要条件.

1.   预备知识
  • 本文约定$\mathbb{R}$+=[0,∞),$\mathbb{N}$为自然数集,Ø为空集.

    定义 1[18]  设二元算子*:[0, 1]×[0, 1]→[0, 1]满足:∀abcd∈[0, 1],

    (a) *对结合律和交换律成立;

    (b) *是连续的;

    (c) a*1=a,∀a∈[0, 1];

    (d) 当acbd时,a*bc*d.

    则称*是连续t-模,常用的连续t-模包括以下3个算子:∀ab∈[0, 1],a*b=aba*b=max{a+b-1,0},a*b=a·b.

    性质 1[9]  设*是连续t-模,

    (i) 若0<r2r1<1,则存在r3∈(0,1),使得r1*r3r2

    (ii) ∀r4∈(0,1),存在r5∈(r4,1),使得r5*r5r4.

    定义 2  设X是一非空集合,*是连续t-模,X上的映射MX2×(0,∞)→(0,1]满足条件:对任意的xyzX

    (M1) ∀t>0,M(xyt)>0;

    (M2) ∀t>0,M(xyt)=1当且仅当x=y

    (M3) ∀t>0,M(xyt)=M(yxt);

    (M4) ∀ts>0,M(xyt)*M(yzs)≤M(xzt+s);

    (M5) M(xy,·):(0,∞) (0,1]是左连续的;

    (M6) $\mathop {\lim }\limits_{t \to \infty } $M(xyt)=1.

    则称(M,*)(简写成M)是X上的模糊度量,称(XM,*)为模糊度量空间.

    注 1  如果将定义2中的(M2),(M5)分别改为:

    (M2)′ M(xyt)=1当且仅当x=y

    (M5)′ M(xy,·)是连续的.

    则(XM,*)为GV模糊度量空间[2].

    若(XM,*)是模糊度量空间,设xXr∈(0,1),t>0,称

    是以x为心,r为半径的开球. 定理1的证明可参见文献[2]中相应结论的证明.

    定理 1  设(XM,*)是模糊度量空间. 如果

    τMX上的第一可数的拓扑,$\left\{ {{B_M}\left( {x, \frac{1}{n}, \frac{1}{n}} \right):n \in \mathbb{N}} \right\}$是点x的可数邻域基.

2.   星伪度量族
  • 本节引入星伪度量族的概念,并给出模糊度量的星伪度量族分解定理.

    定义 3  设X是一非空集合,*是连续t-模,{drr∈(0,1)}是X×X$\mathbb{R}$+中的一族映射. 若对任意的xyzX,都有:

    (SPM1) ∀t>0,存在r∈(0,1),使得dr(xy)≤t

    (SPM2) ∀r∈(0,1),dr(xx)=0;

    (SPM3) ∀r∈(0,1),dr(xy)=dr(yx);

    (SPM4) 对固定的xyX,关于r∈(0,1)的函数dr(xy)是单调递增的;

    (SPM5) 对任意的αβ∈(0,1),dα*β(xz)≤dα(xy)+dβ(yz);

    (SPM6) 若xy,则$\mathop {\sup }\limits_{r \in \left( {0, 1} \right)} $dr(xy)>0.

    则称{drr∈(0,1)}是X上的星伪度量族,称(Xdrr∈(0,1))为星伪度量族空间.

    注 2  当*=∧时,星伪度量族即为伪度量族. 对于一般的连续t-模*,星伪度量族中的元素未必为伪度量,但为方便起见,我们仍称其为星伪度量族.

    为与星伪度量族空间作区分,我们将由X上的一族伪度量{drr∈(0,1)}构成的空间(Xdrr∈(0,1))称为伪度量族空间.

    定理 2  设X是一非空集合,D={drr∈(0,1)}是X上的星伪度量族,对任意的xXn$\mathbb{N}$r1r2,…,rn∈(0,1)和ε>0,

    X存在唯一的拓扑τD,使得对任意的xX

    恰好是x关于τD的邻域基,且τDX上的Hausdorff拓扑.

      前半部分的证明是常规的,这里仅给出τD是Hausdorff拓扑的证明.

    事实上,对任意的不同的点xyX,由条件(SPM6),存在r∈(0,1),使得dr(xy)=ε>0. 由性质1(ii),存在s∈(r,1),使得s*sr,从而Vx$\left( {s;\frac{\varepsilon }{2}} \right)$Vx以及Vy$\left( {s;\frac{\varepsilon }{2}} \right)$Vy. 利用条件(SPM4)和(SPM5)可验证Vx$\left( {s;\frac{\varepsilon }{2}} \right)$Vy$\left( {s;\frac{\varepsilon }{2}} \right)$ =Ø. 因此τD是Hausdorff的.

    引理 1  设(XM,*)为模糊度量空间,xyXr∈(0,1). 则:

    (i) sup{t>0:M(xyt)≤r}=inf{t>0:M(xyt)>r};

    (ii) sup{t>0:M(xyt)<r}=inf{t>0:M(xyt)≥r};

    (iii) inf{t>0:M(xyt)>r}=inf{t>0:M(xyt)≥r}当且仅当M(xy,·)是严格单调增的.

    定理 3  设(XM,*)为模糊度量空间,xyXr∈(0,1). 令

    DM={drr∈(0,1)}是星伪度量族.

      只要证DM满足条件(SPM1)-(SPM6)即可. (SPM2)和(SPM3)是显然的.

    (SPM1):∀xyXt>0,由M(xyt)>0,则存在r0∈(0,1)使得M(xyt)>r0>0. 由(3)式得dr0(xy)≤t.

    (SPM4):任取r1r2∈(0,1),r1r2. 因为M(xy,·)是单调增的,所以

    所以dr1(xy)≥dr2(xy). 因此dr(xy)关于r∈(0,1)是单调增的.

    (SPM5):对任意的xyzXαβ∈(0,1),任取t1dα(xz)及t2dβ(zy). 由(3)式知,存在t1*t2*>0,使得t1*t1t2*t2,且M(xzt1*)≥αM(zyt2*)≥β. 因此

    从而dα*β(xy)≤t1*+t2*t1+t2. 根据t1t1的任意性得dα*β(xy)≤dα(xz)+dβ(zy).

    (SPM6):∀xyXxy,由定义2,存在t0>0使得M(xyt0)<1. 取r0∈(0,1)满足M(xyt0)<r0<1,由引理1得

    因此dr0(xy)≥r0>0,从而$\mathop {\sup }\limits_{r \in \left( {0, 1} \right)} $dr(xy)>0.

    注 3  称上述D={drr∈(0,1)}为由模糊度量M导出的星伪度量族.

    定理 4  设D={drr∈(0,1)}为X上的星伪度量族,对xyXt>0,设

    则(XMD,*)是一个模糊度量空间.

      以下证明MD满足条件(M1)-(M6). (M3)显然成立.

    (M1):对任意的t>0,取0<tt. 由条件(SPM1),存在r0∈(0,1),使得dr0(xy)≤tt. 因此MD(xyt)≥r0>0.

    (M2):令x=y. 由条件(SPM2),对任意的r∈(0,1),t>0,有tdr(xy)=0. 因此

    相反地,假设对任意的t>0,有MD(xyt)=1,则对任意的r∈(0,1),MD(xyt)>r. 由(4)式知,存在1>rr,使得dr(xy)<t. 由条件(SPM4)得dr(xy)<t. 再由t的任意性知dr(xy)=0. 根据r的任意性和条件(SPM6)知x=y.

    (M4):任取xyzXts>0,令MD(xyt)=βMD(yzs)=γ. 对任意的ε>0且ε<min{βγ},由(4)式知,存在r′,r″∈(0,1),使得r′>β-εr″>γ-εdr(xy)<tdr(yz)<s. 因此dβ-ε(xy)<tdγ-ε(yz)<s. 不失一般性,假设βγ,则

    因此MD(xzt+s)≥γ-ε. 由ε的任意性和*算子的连续性可得

    (M5):对任意的xyXt0>0和ε>0,有MD(xyt0)-εMD(xyt0). 由(4)式知,存在r0∈(0,1),使得dr0(xy)<t0r0MD(xyt0)-ε,即MD(xyt0)-r0ε. 当dr0(xy)<tt0时,由(4)式知MD(xyt)≥r0,因此

    也就是说MD(xy,·)在t0处是左连续的. 再由t0的任意性知MD(xy,·)是左连续的.

    (M6):∀xyXt0>0和ε>0,取r0∈(0,1),使得1-r0ε. 由(4)式,当tdr0(xy)时,MD(xyt)≥r0>1-ε. 因此$\mathop {\lim }\limits_{t \to \infty } $MD(xyt)=1.

3.   模糊度量空间中的等距同构
  • 在本节中,我们将研究模糊度量和伪度量族之间的等距同构关系.

    定义 4  设(XM,*)是模糊度量空间,(Xdrr∈(0,1))是星伪度量族空间. 若存在XX上的一一映射Φ,使得∀xyXt>0,∀r∈(0,1),都有

    则称Φ是(XM,*)到(Xdrr∈(0,1))上的等距同构映射,称模糊度量空间(XM,*)等距同构于星伪度量族空间(Xdrr∈(0,1)).

    定理 5  若Φ是从模糊度量空间(XM,*)到星伪度量族空间(Xdrr∈(0,1))上的等距同构映射,则Φ是从(XτM)的到(XτD)的同胚映射,其中τM如(2)式定义,τD是由星伪度量族{drr∈(0,1)}导出的拓扑.

      只需证ΦΦ的逆映射Φ-1都是连续的,只要证明:

    (a) 对任意的xXt>0和r∈(0,1),Φ(BM$\left( {x, 1 - r, \frac{t}{2}} \right)$)⊆VΦ(x)(tr);

    (b) 对任意的xXt>0及r∈(0,1),Φ-1(Vx $\left( {t, 1 - \frac{r}{2}} \right)$)⊆BM(Φ-1(x),rt).

    具体证明过程是常规的.

    推论 1  设(XM,*)为模糊度量空间,DM={drr∈(0,1)}为由M生成的星伪度量族,则由M诱导的拓扑τM与其对应的由星伪度量族所诱导的拓扑τDM是一致的.

    定义 5  设(XM,*)和(XM,*)是两个模糊度量空间,若存在XX上的一一映射ψ,使得∀xyXt>0,都有M(xyt)=M(ψ(x),ψ(y),t),则称ψ是(XM,*)到(XM,*)上的等距同构映射,称模糊度量空间(XM,*)等距同构于模糊度量空间(XM,*).

    定理 6  若模糊度量空间(XM,*)由星伪度量族空间(Xdrr∈(0,1))导出,模糊度量空间(XM,*)等距同构于(Xdrr∈(0,1)),那么(XM,*)也等距同构于(XM,*).

      设Φ是模糊度量空间(XM,*)到伪度量族空间(Xdrr∈(0,1))的等距同构映射,那么,∀xyXr∈(0,1),都有(5)式成立. 由(4)式知:∀t>0,有

    dr(Φ(x),Φ(y))<t,则由(5)式可知,存在0<tt,使得M(xyt)≥r. 再由M(xy,·)单调增可得M(xyt)≥r. 所以

    对任意给定的0<rM(xyt),由M(xy,·)的左连续性知,存在0<tt,使得M(xyt)>r. 由(5)式得dr(Φ(x),Φ(y))≤tt. 于是

    r的任意性知

    由(7)式和(8)式知

    由(6)式知

    从而(XM,*)等距同构于(XM,*).

    推论 2  设(XM,*)为模糊度量空间,DM={drr∈(0,1)}为由M生成的星伪度量族,MDM为由DM={drr∈(0,1)}导出的模糊度量,则(XM,*)等距同构于(XMDM,*). 因此,由MMDM导出的拓扑是一致的.

    定理 7  设(XM,*)为模糊度量空间,若(XM,*)满足条件:对任意的xyzXst>0,有

    则(XM,*)等距同构于某个伪度量族空间(Xdrr∈(0,1)).

      取X=XΦ(x)=x(∀xX). 设dr(xy)由(3)式定义,则由定理3知dr(xy)为X上的分离的星伪度量族.

    下证{drr∈(0,1)}为伪度量族. 只需证每个dr满足三角不等式. 为此任取xyzXr∈(0,1). 对任意的t1dr(xz),t2dr(zy),由(3)式,存在t1*t2*>0,使得t1*t1t2*t2M(xzt1*)≥rM(zyt2*)≥r. 由(9)式知

    于是由(3)式知

    再由t1t2的任意性得

    最后,由dr(xy)的定义即知(XM,*)与(Xdrr∈(0,1))等距同构.

    推论 3  每个模糊度量(XM,∧)都可以被分解成X上的一族伪度量.

    容易证明,若M(xy,·)是连续的,则定理7的逆定理也成立,即:

    定理 8  设模糊度量空间(XM,*)满足条件:对任意的xyX,函数M(xy,·)连续. 若(XM,*)等距同构于某个伪度量族空间(Xdrr∈(0,1)),则不等式(9)成立.

Reference (18)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return